A camera autofocus method based on visual saliency

Abstract

The invention discloses a camera automatic focusing method based on visual saliency, and its main ideas are as follows: determine the image matrix I, averagely divide the image matrix I into α·β blocks, and calculate the Lab of the image matrix I In the color space, the color a‑channel corresponds to the channel image I a‑channel And the color b channel in the Lab color space of the image matrix I corresponds to the channel image I b‑channel , calculate the block P at the horizontal number a and the vertical number b in the image matrix I a,b The total significance weight W of a,b ;a=1,2,...,α,b=1,2,...,β; Calculate the block P at the horizontal number a and the vertical number b in the image matrix I a,b The original contrast value C a,b ; Calculate the block P at the horizontal number a and the vertical number b in the image matrix I a,b The significance weighted contrast value C`` a,b ; a=1,2,...,α,b=1,2,...,β; and then obtain the saliency weighted contrast value of each block in the image matrix I, recorded as α×β saliency in the image matrix I Weighting the contrast value; then determining the focus area; calculating the contrast value C of the focus area, and obtaining the best focus position according to the contrast value C of the focus area, and then completing the focus.

Description

technical field [0001] The invention belongs to the technical field of television cameras, in particular to a camera auto-focus method based on visual salience, which is suitable for auto-focus area selection and contrast focus. Background technique [0002] Visual salience is to extract an interesting subset of images before performing advanced visual processing to reduce the complexity of processing. In actual photography, there are usually salient things in the scene, which can be selected by selecting salient areas as The focus area can optimize the focus effect; contrast focus, also called contrast focus, is an algorithm that uses the contrast of objects in the image to focus. After imaging, the next step is determined according to the imaging effect. Contrast focus is widely used in camera systems, but the current focus algorithm The automatic area selection function of the camera is relatively weak, resulting in the selection of the focus area is not smart enough, and...

AI Technical Summary

Problems solved by technology

[0002] Visual salience is to extract an interesting subset of images before performing advanced visual processing to reduce the complexity of processing. In actual photography, there are usually salient things in the scene, which can be selected by selecting salient areas as The focus area can optimize the focus effect; contrast focus, also called contrast focus, is an algorithm that uses the contrast of objects in the image to focus. After imaging, the next step is determined according to the imaging effect. Contrast focus is widely used in camera systems, but the current focus algorithm The automatic area selection function of the software is relatively weak, resulting in the selection of the focus area is not smart enough, and the focus effect is not satisfactory

Images